INDEX

Explanations

mentions of family relationships, particularly the word "cousin"

oai_token-act-pair · gpt-3.5-turbo

references to familial relationships, specifically terms related to cousins

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

inth

-0.84

hner

-0.79

inen

-0.75

inem

-0.74

anwhile

-0.73

yss

-0.70

ights

-0.70

overe

-0.69

Ö¼

-0.69

arching

-0.66

POSITIVE LOGITS

 cousin

0.93

 cousins

0.92

 aunt

0.89

 uncle

0.89

 nephew

0.87

 niece

0.86

hesis

0.85

hood

0.79

 incest

0.73

uncle

0.71

Activations Density 0.010%

mentions of family relationships, particularly the word "cousin"

references to familial relationships, specifically terms related to cousins

No Comments

No Known Activations