INDEX

Explanations

mentions of the word "rat" and related words

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "rat."

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

qua

-0.73

PLIED

-0.72

OURCE

-0.70

Ļ½

-0.69

»Ĵ

-0.68

RECT

-0.64

ONT

-0.61

«ĺ

-0.60

âĢ¢âĢ¢

-0.60

 whence

-0.59

POSITIVE LOGITS

chet

1.29

rat

1.10

Rat

0.85

ented

0.84

bats

0.82

oxin

0.82

 Rats

0.80

imilation

0.79

holes

0.79

cat

0.79

Activations Density 0.006%

mentions of the word "rat" and related words

mentions of the word "rat."

No Comments

No Known Activations