INDEX

Explanations

words related to communication such as "caller" and "listener."

oai_token-act-pair · gpt-3.5-turbo

words related to someone receiving or engaging in communication, like "caller" or "listener"

oai_token-act-pair · gpt-4-turbo

references to individuals involved in a call or communication context, specifically callers and listeners

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Pradesh

-0.74

rity

-0.73

 Centers

-0.70

 Catholicism

-0.68

 Crimes

-0.65

ceed

-0.65

rams

-0.64

 Formation

-0.63

 Paint

-0.62

 Bhar

-0.60

POSITIVE LOGITS

 beware

1.05

iquid

0.85

 interviewer

0.81

 Beware

0.80

 misunderstood

0.80

 extraord

0.80

 listeners

0.79

esses

0.78

backer

0.77

 reviewer

0.76

Activations Density 0.043%

words related to communication such as "caller" and "listener."

words related to someone receiving or engaging in communication, like "caller" or "listener"

references to individuals involved in a call or communication context, specifically callers and listeners

No Comments

No Known Activations