INDEX

Explanations

proper nouns, likely names of people and places

oai_token-act-pair · gpt-3.5-turbo

characters or letters of the alphabet

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

urat

-0.59

 Klux

-0.59

Measure

-0.57

ELF

-0.56

orget

-0.56

APD

-0.55

 subtract

-0.54

oway

-0.54

 sergeant

-0.54

 contacting

-0.53

POSITIVE LOGITS

uala

0.83

icz

0.80

ito

0.76

ildo

0.73

andro

0.73

icular

0.73

theless

0.69

ius

0.69

ulum

0.68

tics

0.68

Activations Density 0.288%

proper nouns, likely names of people and places

characters or letters of the alphabet

No Comments

No Known Activations