INDEX

Explanations

mentions of bounties or rewards for specific actions or information

oai_token-act-pair · gpt-3.5-turbo

references to "bounty" in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Edinburgh

-0.74

 Chart

-0.74

 Olson

-0.73

ÅŁ

-0.68

 Rooms

-0.67

cu

-0.67

 Cancel

-0.66

 Clin

-0.66

apses

-0.66

ovic

-0.66

POSITIVE LOGITS

 bounty

4.28

 bount

3.23

 Bounty

3.20

ounty

2.08

 reward

1.36

 ransom

1.30

Reward

1.29

 Reward

1.21

 rewards

1.16

 karma

1.12

Activations Density 0.019%

mentions of bounties or rewards for specific actions or information

references to "bounty" in various contexts

No Comments

No Known Activations