INDEX

Explanations

descriptions of items or scenarios being complete with additional features

oai_token-act-pair · gpt-3.5-turbo

references to completeness or additional components in descriptions

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

thren

-0.82

sts

-0.80

ãģ®éŃĶ

-0.75

Asia

-0.70

ajor

-0.66

nai

-0.66

ungle

-0.66

station

-0.66

formation

-0.64

bane

-0.64

POSITIVE LOGITS

 regard

1.03

draw

0.89

 impunity

0.87

 apologies

0.77

 respect

0.77

 caveats

0.77

 jewels

0.76

 regards

0.76

£ı

0.76

 mockery

0.75

Activations Density 0.082%

descriptions of items or scenarios being complete with additional features

references to completeness or additional components in descriptions

No Comments

No Known Activations