INDEX

Explanations

code-related terms and specific code snippets

oai_token-act-pair · gpt-3.5-turbo

references to "code" and related terms in programming discussions

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

 Flavoring

-0.96

iatus

-0.83

iversal

-0.79

 Democr

-0.76

iscopal

-0.71

 Wynne

-0.70

 Chero

-0.68

izo

-0.66

 disadvant

-0.66

cffff

-0.66

POSITIVE LOGITS

 snippet

1.34

base

1.11

 snippets

1.08

code

0.98

 code

0.97

unit

0.87

 snipp

0.84

velop

0.83

works

0.83

 repository

0.83

Activations Density 0.019%

code-related terms and specific code snippets

references to "code" and related terms in programming discussions

No Comments

No Known Activations