INDEX

Explanations

sequences of random characters and symbols

oai_token-act-pair · gpt-3.5-turbo

Japanese characters and 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

 disadvant

-0.92

raints

-0.89

pheus

-0.83

 manif

-0.82

schild

-0.80

 mathemat

-0.76

undai

-0.76

 constitu

-0.74

 subord

-0.73

 Seym

-0.73

POSITIVE LOGITS

ħ

0.94

 âĢº

0.87

ļ

0.87

ï¸ı

0.84

İ

0.83

ı

0.83

Ķ

0.82

ĺ

0.82

Ī

0.82

®

0.82

Activations Density 0.056%

sequences of random characters and symbols

Japanese characters and symbols

No Comments

No Known Activations