INDEX

Explanations

references to receiving a bonus

oai_token-act-pair · gpt-3.5-turbo

references to bonuses and rewards

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ŀ

-0.76

ells

-0.74

brew

-0.74

URRENT

-0.73

 Wiley

-0.72

doi

-0.71

è£

-0.70

 Nicotine

-0.70

 Conflict

-0.69

ALE

-0.69

POSITIVE LOGITS

 bonus

1.38

 bonuses

1.33

 redundancy

0.98

 depreciation

0.92

 parachute

0.89

 incentive

0.84

 accrued

0.84

 accol

0.84

 sever

0.84

Bonus

0.83

Activations Density 0.009%

references to receiving a bonus

references to bonuses and rewards

No Comments

No Known Activations