INDEX

Explanations

years, specifically 2013

oai_token-act-pair · gpt-3.5-turbo

occurrences of the year 2013

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

flo

-0.82

srf

-0.76

warts

-0.74

ntil

-0.71

sembly

-0.70

htaking

-0.69

 ship

-0.69

kef

-0.69

phy

-0.68

boats

-0.68

POSITIVE LOGITS

å¹

0.95

 onwards

0.81

-'

0.80

rd

0.78

0.76

0.74

aretz

0.72

 onward

0.71

 edition

0.70

Activations Density 0.021%

years, specifically 2013

occurrences of the year 2013

No Comments

No Known Activations