INDEX

Explanations

proper nouns of individuals prevalent in specific events or contexts

oai_token-act-pair · gpt-3.5-turbo

references to individuals or entities being mentioned together

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

Trend

-0.73

veyard

-0.70

artz

-0.68

blem

-0.64

slaught

-0.64

 strang

-0.64

 Warden

-0.64

 Barrier

-0.62

hibition

-0.62

anyl

-0.62

POSITIVE LOGITS

 vying

0.81

 equally

0.79

 congratulated

0.73

edged

0.72

equal

0.71

 halves

0.71

 respectively

0.70

 sides

0.70

ocating

0.67

 separately

0.67

Activations Density 0.092%

proper nouns of individuals prevalent in specific events or contexts

references to individuals or entities being mentioned together

No Comments

No Known Activations