INDEX

Explanations

phrases indicating familiarity or experience with a particular subject

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate familiarity or exceptionality in a context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ahime

-0.88

 buggy

-0.66

edia

-0.62

 Hitch

-0.61

irez

-0.61

 Software

-0.60

WI

-0.59

 Restoration

-0.58

 Forestry

-0.58

 Unified

-0.56

POSITIVE LOGITS

 whatsoever

1.10

achine

0.72

aunts

0.69

bleacher

0.69

 riches

0.65

Reviewer

0.65

uge

0.63

tips

0.63

gon

0.63

REDACTED

0.63

Activations Density 0.064%

phrases indicating familiarity or experience with a particular subject

phrases that indicate familiarity or exceptionality in a context

No Comments

No Known Activations