INDEX

Explanations

phrases related to question and answer sessions (Q&A)

oai_token-act-pair · gpt-3.5-turbo

patterns related to question-and-answer formats

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ngth

-0.99

 crest

-0.73

ãĤ¼ãĤ¦ãĤ¹

-0.71

pread

-0.68

 Plum

-0.66

ãĥ´ãĤ¡

-0.65

ãĥ¢

-0.65

ufact

-0.65

olesterol

-0.63

ãĤ¨ãĥ«

-0.63

POSITIVE LOGITS

spir

0.72

Tel

0.65

border

0.64

monary

0.63

abus

0.63

ĪĴ

0.62

Dialog

0.62

aq

0.62

Desk

0.62

Country

0.61

Activations Density 0.093%

phrases related to question and answer sessions (Q&A)

patterns related to question-and-answer formats

No Comments

No Known Activations