INDEX

Explanations

values representing ratios between different entities

oai_token-act-pair · gpt-3.5-turbo

statistical comparisons expressed in ratio format

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

gotten

-0.74

pocket

-0.70

details

-0.66

dream

-0.64

sylvania

-0.64

uggage

-0.64

awar

-0.63

ridges

-0.62

REAM

-0.61

asper

-0.61

POSITIVE LOGITS

xual

1.09

 ratios

0.92

yip

0.91

 ratio

0.83

nces

0.80

ically

0.79

ical

0.78

 mismatch

0.74

ally

0.74

SHIP

0.70

Activations Density 0.027%

values representing ratios between different entities

statistical comparisons expressed in ratio format

No Comments

No Known Activations