INDEX

Explanations

quantitative terms related to numbers or figures

oai_token-act-pair · gpt-3.5-turbo

quantifiable measurements or totals in text

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

aunts

-0.79

tics

-0.78

laws

-0.77

 Guys

-0.76

ooks

-0.74

birds

-0.74

stones

-0.72

pins

-0.71

acies

-0.71

assies

-0.71

POSITIVE LOGITS

 amount

0.96

 number

0.84

0.78

 ratio

0.78

©¶æ¥µ

0.75

 percentage

0.75

 degree

0.74

ãĥĩãĤ£

0.73

 coefficient

0.72

ccording

0.72

Activations Density 0.129%

quantitative terms related to numbers or figures

quantifiable measurements or totals in text

No Comments

No Known Activations