INDEX

Explanations

methods or techniques within a text

oai_token-act-pair · gpt-3.5-turbo

references to specific methods or techniques

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

igi

-0.64

 Barkley

-0.62

minster

-0.62

 Falcons

-0.61

enda

-0.60

 Sons

-0.59

orb

-0.59

 challeng

-0.59

vation

-0.58

 Childhood

-0.58

POSITIVE LOGITS

ologies

1.32

ologically

1.11

ologic

1.00

OLOGY

0.94

ology

0.92

ical

0.92

 method

0.84

ically

0.83

idable

0.81

sonian

0.79

Activations Density 0.020%

methods or techniques within a text

references to specific methods or techniques

No Comments

No Known Activations