INDEX

Explanations

words related to additional rewards or incentives

oai_token-act-pair · gpt-3.5-turbo

references to bonus incentives or rewards

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

tics

-0.88

ãĤ¢ãĥ«

-0.81

scape

-0.81

turn

-0.76

ichen

-0.74

URES

-0.73

tic

-0.73

lÃ©

-0.72

Ö¼

-0.72

INK

-0.71

POSITIVE LOGITS

 bonus

1.15

 bonuses

1.00

 depreciation

0.92

 accrued

0.86

 parach

0.78

 Bonus

0.78

 extras

0.77

Reviewer

0.77

 parachute

0.76

 indemn

0.75

Activations Density 0.014%

words related to additional rewards or incentives

references to bonus incentives or rewards

No Comments

No Known Activations