INDEX

Explanations

names of people and statements they made in news articles

oai_token-act-pair · gpt-3.5-turbo

instances of reported speech or statements made by individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

à¦

-0.86

ptives

-0.84

folios

-0.84

otin

-0.83

Newsletter

-0.78

arest

-0.74

ãĥĸ

-0.74

izont

-0.72

Translation

-0.72

ption

-0.71

POSITIVE LOGITS

 goodbye

0.90

 afterward

0.88

 anecd

0.88

 Hancock

0.82

doms

0.82

 sergeant

0.76

 examiner

0.73

 superintendent

0.73

 afterwards

0.73

 Speedway

0.72

Activations Density 0.215%

names of people and statements they made in news articles

instances of reported speech or statements made by individuals

No Comments

No Known Activations