INDEX

Explanations

names of locations and organizations

oai_token-act-pair · gpt-3.5-turbo

mentions of the D.C. Circuit Court

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

onyms

-0.76

VALUE

-0.69

onym

-0.69

ventus

-0.69

ä

-0.68

ãĥı

-0.65

angible

-0.65

CLASS

-0.65

ãĥ«

-0.64

ãĥ¤

-0.64

POSITIVE LOGITS

 Fields

0.76

 Speedway

0.71

 Retrieved

0.70

 Agency

0.66

 Clim

0.66

 Pass

0.66

 Residents

0.65

 Border

0.65

 Liberties

0.65

 Torn

0.65

Activations Density 0.048%

names of locations and organizations

mentions of the D.C. Circuit Court

No Comments

No Known Activations