INDEX

Explanations

terms related to evaluation and analysis

oai_token-act-pair · gpt-3.5-turbo

concepts related to analysis and evaluation of limitations or deficiencies

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

enegger

-0.67

ä»£

-0.65

aturdays

-0.60

eday

-0.60

ÃĥÃĤ

-0.59

bye

-0.59

AAAAAAAA

-0.57

daq

-0.57

nesday

-0.56

retty

-0.56

POSITIVE LOGITS

 harmful

0.55

Â¶

0.53

 certain

0.50

 Entered

0.48

 behaviors

0.48

 âĢº

0.47

 populations

0.47

 microbiota

0.47

 harms

0.46

ÂŃ

0.46

Activations Density 1.596%

terms related to evaluation and analysis

concepts related to analysis and evaluation of limitations or deficiencies

No Comments

No Known Activations