INDEX

Explanations

terms related to specific specialized areas or domains

oai_token-act-pair · gpt-3.5-turbo

references to specific areas or fields of study

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

pter

-0.75

 Masquerade

-0.74

itous

-0.72

[&

-0.66

arus

-0.66

 Sons

-0.66

icious

-0.65

inn

-0.64

 ducks

-0.62

 procession

-0.62

POSITIVE LOGITS

 areas

0.89

 area

0.84

 proble

0.81

icals

0.78

manship

0.77

utics

0.68

 disciplines

0.65

 expertise

0.65

 matters

0.65

folios

0.65

Activations Density 0.046%

terms related to specific specialized areas or domains

references to specific areas or fields of study

No Comments

No Known Activations