INDEX

Explanations

important or noteworthy points in a story

oai_token-act-pair · gpt-3.5-turbo

keywords related to news summaries and highlights

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

 nurs

-0.79

ogly

-0.76

haps

-0.72

oplan

-0.72

unci

-0.71

Magikarp

-0.71

rag

-0.69

lia

-0.67

nib

-0.67

ooked

-0.66

POSITIVE LOGITS

 WATCHED

0.86

 Highlights

0.84

 Kaine

0.83

 VIDEOS

0.78

 Legislation

0.76

ï

0.74

 Volunte

0.74

 Timeline

0.73

 Decoder

0.72

 Researchers

0.71

Activations Density 0.015%

important or noteworthy points in a story

keywords related to news summaries and highlights

No Comments

No Known Activations