INDEX

Explanations

occurrences of the word "no"

oai_token-act-pair · gpt-3.5-turbo

negative assertions

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

UC

-0.62

 Sting

-0.60

Spy

-0.58

 gentleman

-0.58

 operative

-0.58

TextColor

-0.56

 Himself

-0.56

 Offic

-0.55

 Judges

-0.55

 Baal

-0.54

POSITIVE LOGITS

ugu

0.78

externalActionCode

0.74

hetti

0.74

ickr

0.71

cffff

0.71

 Compensation

0.70

rontal

0.70

sonian

0.68

iameter

0.65

ramid

0.65

Activations Density 0.000%

occurrences of the word "no"

negative assertions

No Comments

No Known Activations