INDEX

Explanations

words related to taking action or making decisions

oai_token-act-pair · gpt-3.5-turbo

variations of the word "cares."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 polymer

-0.69

UE

-0.65

Ill

-0.62

æĸ¹

-0.61

Assembly

-0.61

IB

-0.60

 ribbon

-0.59

 observable

-0.59

 backbone

-0.58

IME

-0.58

POSITIVE LOGITS

ares

1.40

aurus

0.93

coni

0.88

 Franch

0.87

tsky

0.85

idon

0.77

paren

0.76

pread

0.74

jon

0.74

bian

0.73

Activations Density 0.005%

words related to taking action or making decisions

variations of the word "cares."

No Comments

No Known Activations