INDEX

Explanations

words related to legislative matters or specific legal entities, specifically focusing on the term "AB."

oai_token-act-pair · gpt-3.5-turbo

mentions of specific legislative bills or identifiers

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Fargo

-0.72

 Episcopal

-0.71

andum

-0.71

 Tycoon

-0.71

 Daytona

-0.68

 Hurricane

-0.68

 Cannes

-0.67

 Revival

-0.67

 Seas

-0.66

 Easter

-0.65

POSITIVE LOGITS

stract

1.08

raham

0.93

omination

0.90

AB

0.88

bey

0.88

RAM

0.85

AB

0.83

isoft

0.82

sembly

0.81

 challeng

0.81

Activations Density 0.002%

words related to legislative matters or specific legal entities, specifically focusing on the term "AB."

mentions of specific legislative bills or identifiers

No Comments

No Known Activations