INDEX

Explanations

non-English characters and phrases

oai_token-act-pair · gpt-3.5-turbo

non-English characters or symbols

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

ktop

-0.78

 Collider

-0.75

 sensit

-0.70

heit

-0.69

enegger

-0.67

 Bentley

-0.63

 manif

-0.62

sein

-0.61

 puzz

-0.61

 sensation

-0.60

POSITIVE LOGITS

ĺ

0.95

ļ

0.94

ÑĢ

0.91

ł

0.89

ĳ

0.88

į

0.88

ª

0.86

ħ

0.85

¹

0.85

Ĺ

0.85

Activations Density 0.048%

non-English characters and phrases

non-English characters or symbols

No Comments

No Known Activations