INDEX

Explanations

words related to assessment or judgment

oai_token-act-pair · gpt-3.5-turbo

references to evaluations and assessments

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

fold

-0.85

hold

-0.81

fall

-0.80

blue

-0.78

birth

-0.77

seed

-0.72

feeding

-0.70

tile

-0.69

href

-0.69

fast

-0.68

POSITIVE LOGITS

 evaluates

0.99

 evaluations

0.99

uations

0.97

 evaluation

0.91

aution

0.86

 appraisal

0.83

uate

0.80

 evaluate

0.79

 evaluating

0.78

ments

0.78

Activations Density 0.019%

words related to assessment or judgment

references to evaluations and assessments

No Comments

No Known Activations