INDEX

Explanations

mentions of open letters or discussions

oai_token-act-pair · gpt-3.5-turbo

references to openness and transparency in communication or society

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

iage

-0.73

ibur

-0.66

 Commando

-0.64

CTV

-0.63

doms

-0.62

********************************

-0.61

 Marketable

-0.60

yz

-0.59

dos

-0.58

 Beasts

-0.58

POSITIVE LOGITS

 minded

0.89

hearted

0.83

ended

0.82

 source

0.81

street

0.81

office

0.76

 doors

0.75

 eyed

0.74

 door

0.73

ource

0.73

Activations Density 0.045%

mentions of open letters or discussions

references to openness and transparency in communication or society

No Comments

No Known Activations