INDEX

Explanations

terms related to evaluations or assessments

oai_token-act-pair · gpt-3.5-turbo

instances of the word "evaluation" and its variations

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

hold

-0.88

fall

-0.82

ship

-0.73

rough

-0.69

fold

-0.69

fu

-0.68

Cry

-0.66

ker

-0.66

Hat

-0.66

lines

-0.66

POSITIVE LOGITS

 evaluations

1.25

 evaluates

1.21

 evaluation

1.19

uations

1.08

 evaluate

0.97

 Eval

0.97

ãħĭãħĭ

0.93

ADRA

0.92

 eval

0.92

 evaluating

0.92

Activations Density 0.011%

terms related to evaluations or assessments

instances of the word "evaluation" and its variations

No Comments

No Known Activations