INDEX

Explanations

specific legal or regulatory provisions within text

oai_token-act-pair · gpt-3.5-turbo

references to legislative provisions and related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ampa

-0.72

issance

-0.70

 Carbuncle

-0.68

photo

-0.68

Pie

-0.66

sburgh

-0.65

Textures

-0.63

roy

-0.61

unit

-0.61

image

-0.60

POSITIVE LOGITS

 prohibiting

1.13

 inserted

1.06

witz

1.04

 stip

1.02

 authorizing

1.00

ality

1.00

 repealing

0.98

 barring

0.93

 exempted

0.93

 repealed

0.92

Activations Density 0.098%

specific legal or regulatory provisions within text

references to legislative provisions and related terms

No Comments

No Known Activations