INDEX

Explanations

Twitter usernames and links

oai_token-act-pair · gpt-3.5-turbo

mentions or references to Twitter

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

 Sphere

-0.73

 Suff

-0.70

 Dahl

-0.67

 Shack

-0.66

 Britann

-0.63

 Wonderland

-0.63

Interstitial

-0.63

uilding

-0.62

 Forth

-0.61

 Rhod

-0.61

POSITIVE LOGITS

ileaks

0.84

twitter

0.81

sers

0.76

username

0.76

ipe

0.75

/#

0.75

"@

0.74

 hashtag

0.74

 feed

0.71

sburg

0.71

Activations Density 0.014%

Twitter usernames and links

mentions or references to Twitter

No Comments

No Known Activations