INDEX

Explanations

mentions of politicians

oai_token-act-pair · gpt-3.5-turbo

references to politicians

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

urious

-0.70

uras

-0.69

uran

-0.69

actory

-0.66

 Condition

-0.66

lights

-0.66

âĶĢâĶĢâĶĢâĶĢ

-0.65

ventory

-0.64

zig

-0.64

gged

-0.64

POSITIVE LOGITS

clinton

1.06

hips

0.85

 appoint

0.81

icians

0.80

 correctness

0.79

 impe

0.78

hip

0.77

 elected

0.73

ervatives

0.73

woman

0.70

Activations Density 0.024%

mentions of politicians

references to politicians

No Comments

No Known Activations