INDEX

Explanations

words related to opinions or emotions expressed in text

oai_token-act-pair · gpt-3.5-turbo

references to sentiments or emotional states

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

ctic

-0.70

DERR

-0.70

ARDS

-0.70

arters

-0.68

ummies

-0.68

ARD

-0.66

 Mamm

-0.66

TRI

-0.63

 Shut

-0.62

ÄŁ

-0.62

POSITIVE LOGITS

ality

1.16

 sentiments

1.13

 sentiment

1.03

ally

0.83

 expressed

0.80

ual

0.80

TextColor

0.79

igue

0.78

naire

0.78

uated

0.77

Activations Density 0.013%

words related to opinions or emotions expressed in text

references to sentiments or emotional states

No Comments

No Known Activations