INDEX

Explanations

words related to automatic actions or processes

oai_token-act-pair · gpt-3.5-turbo

terms related to automated processes or functionalities

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

ĸļ

-0.98

ador

-0.80

ergus

-0.79

,,,,

-0.75

 Emin

-0.74

rug

-0.73

 Mouth

-0.72

elia

-0.72

iddles

-0.71

=-=-=-=-=-=-=-=-

-0.70

POSITIVE LOGITS

 populate

0.89

 assume

0.87

 detects

0.86

 migrate

0.83

 detect

0.83

 regenerate

0.81

 aspir

0.79

 induct

0.79

 confir

0.79

 revert

0.78

Activations Density 0.016%

words related to automatic actions or processes

terms related to automated processes or functionalities

No Comments

No Known Activations