INDEX

Explanations

names of news agencies or media platforms

oai_token-act-pair · gpt-3.5-turbo

references to news agencies or media-related identifiers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

liga

-0.69

-0.65

ŃĶ

-0.62

Magikarp

-0.59

 colleg

-0.58

course

-0.56

anian

-0.56

 hitherto

-0.54

ãģ®é

-0.54

 unnecess

-0.53

POSITIVE LOGITS

 Listen

0.82

↵

0.69

 Photo

0.65

 ÃĹ

0.65

toggle

0.64

Buy

0.64

 foreground

0.64

argo

0.62

Enlarge

0.62

photo

0.61

Activations Density 0.017%

names of news agencies or media platforms

references to news agencies or media-related identifiers

No Comments

No Known Activations