INDEX

Explanations

words related to legal proceedings and names

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 bald

-0.76

 Slaughter

-0.75

 Slay

-0.75

 Higgins

-0.74

NetMessage

-0.73

rences

-0.69

CHAR

-0.69

-0.66

 Seat

-0.66

 Chambers

-0.65

POSITIVE LOGITS

1.35

uno

1.28

os

1.11

oing

1.10

emo

1.09

ado

1.09

ino

1.08

lo

1.07

uo

1.06

ro

1.05

Activations Density 0.212%

words related to legal proceedings and names

No Comments

No Known Activations