INDEX

Explanations

events that have occurred multiple times

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "times" and its variations in numerical form

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

rera

-0.67

rats

-0.66

urated

-0.65

CHAT

-0.65

 Redemption

-0.64

league

-0.62

aceous

-0.62

andr

-0.61

ges

-0.61

IELD

-0.60

POSITIVE LOGITS

 consecut

1.07

 apiece

0.92

 throughout

0.86

 before

0.81

 during

0.79

cale

0.74

 already

0.72

per

0.66

points

0.65

online

0.64

Activations Density 0.062%

events that have occurred multiple times

occurrences of the word "times" and its variations in numerical form

No Comments

No Known Activations