INDEX

Explanations

terms related to legal or administrative actions

oai_token-act-pair · gpt-3.5-turbo

terms related to action or legal proceedings

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

¶

-0.69

©¶æ

-0.69

ZI

-0.67

yss

-0.66

 whisk

-0.66

 sliced

-0.64

IX

-0.63

 mortg

-0.62

Veg

-0.60

Ń·

-0.60

POSITIVE LOGITS

action

1.31

actions

1.16

terday

0.90

ivation

0.88

naires

0.86

naire

0.85

ocracy

0.83

ACTION

0.83

olution

0.83

umerable

0.79

Activations Density 0.009%

terms related to legal or administrative actions

terms related to action or legal proceedings

No Comments

No Known Activations