INDEX

Explanations

phrases related to rewards and achievements

oai_token-act-pair · gpt-3.5-turbo

terms related to rewards and incentives

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

abad

-0.76

enium

-0.73

 Ange

-0.72

abases

-0.71

Osw

-0.69

 Extrem

-0.68

 Sloven

-0.68

cling

-0.67

obiles

-0.66

ahime

-0.65

POSITIVE LOGITS

 reward

1.15

 rewards

1.07

 rewarded

1.03

Reward

0.89

 rewarding

0.86

 Reward

0.83

 recipients

0.79

 incentive

0.79

 awarded

0.78

 payoff

0.77

Activations Density 0.024%

phrases related to rewards and achievements

terms related to rewards and incentives

No Comments

No Known Activations