INDEX

Explanations

words related to legal and punishment contexts

oai_token-act-pair · gpt-3.5-turbo

words related to regulations or rules

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 registers

-0.62

lod

-0.61

 flap

-0.58

 latch

-0.58

bra

-0.58

 baked

-0.56

 heaven

-0.55

 phenomen

-0.55

 mount

-0.54

Bro

-0.54

POSITIVE LOGITS

ution

4.28

utions

3.70

UTION

2.40

utive

2.04

uting

1.98

uted

1.89

utes

1.70

ute

1.70

UTE

1.49

utable

1.41

Activations Density 0.012%

words related to legal and punishment contexts

words related to regulations or rules

No Comments

No Known Activations