INDEX

Explanations

words related to insults and derogatory language

oai_token-act-pair · gpt-3.5-turbo

references to insults and derogatory remarks

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

negie

-0.73

ccording

-0.72

ullivan

-0.70

ucket

-0.68

illon

-0.68

ilver

-0.68

ail

-0.67

arten

-0.66

angler

-0.65

ooth

-0.65

POSITIVE LOGITS

ingly

1.02

 insult

0.99

ãĤ¹ãĥĪ

0.89

 insulted

0.86

 insulting

0.83

 insults

0.83

 prejudice

0.77

 disrespect

0.76

 ridicule

0.75

 hurled

0.74

Activations Density 0.022%

words related to insults and derogatory language

references to insults and derogatory remarks

No Comments

No Known Activations