INDEX

Explanations

years

oai_token-act-pair · gpt-3.5-turbo

numerical references to specific years

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ledged

-0.65

illet

-0.61

ndra

-0.59

edge

-0.59

Edge

-0.58

Tu

-0.56

 marqu

-0.55

 hungry

-0.54

 distingu

-0.54

 lightsaber

-0.54

POSITIVE LOGITS

-'

0.95

å¹

0.87

ĸļ

0.84

ãĥŁ

0.77

 onwards

0.76

BCE

0.67

 onward

0.65

–

0.63

é¾įå¥ĳå£«

0.62

chev

0.62

Activations Density 0.067%

years

numerical references to specific years

No Comments

No Known Activations