INDEX

Explanations

Latin characters adjacent to each other

oai_token-act-pair · gpt-3.5-turbo

instances of critical or misleading information

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

oun

-0.76

hement

-0.72

xual

-0.71

 destro

-0.64

itsu

-0.63

 Leilan

-0.63

uto

-0.63

owicz

-0.62

 honoured

-0.62

tsky

-0.62

POSITIVE LOGITS

Â¶

0.85

database

0.84

Â·

0.78

ccording

0.78

é¾

0.76

ÂłÂłÂłÂłÂłÂłÂłÂł

0.75

ÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂł

0.75

natureconservancy

0.73

yet

0.71

JUST

0.71

Activations Density 0.252%

Latin characters adjacent to each other

instances of critical or misleading information

No Comments

No Known Activations