INDEX

Explanations

instances of objects 'coming in'

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "in."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tein

-0.65

aldi

-0.62

 Grab

-0.60

âĢ¢âĢ¢

-0.60

cas

-0.60

IAS

-0.59

DOM

-0.59

ilus

-0.58

 rede

-0.58

FUL

-0.57

POSITIVE LOGITS

 handy

1.13

escap

0.84

between

0.76

 accordance

0.75

offensive

0.73

 somew

0.73

 increments

0.71

versely

0.70

 lieu

0.69

abwe

0.69

Activations Density 0.057%

instances of objects 'coming in'

mentions of the word "in."

No Comments

No Known Activations