INDEX

Explanations

words related to tables

oai_token-act-pair · gpt-3.5-turbo

references to tables of contents or lists in documents

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ovich

-0.72

rily

-0.69

 chancellor

-0.66

 Directorate

-0.66

vernment

-0.65

imal

-0.64

ibly

-0.61

atorium

-0.61

adobe

-0.61

 Enhancement

-0.60

POSITIVE LOGITS

cloth

1.56

au

1.19

aux

1.13

poons

1.05

top

1.02

poon

0.97

tops

0.96

aus

0.92

 manners

0.90

 tennis

0.86

Activations Density 0.038%

words related to tables

references to tables of contents or lists in documents

No Comments

No Known Activations