INDEX

Explanations

words related to a significant and often sudden change

oai_token-act-pair · gpt-3.5-turbo

terms indicating significant changes or transformations

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

tein

-0.99

sburgh

-0.82

imir

-0.75

emis

-0.74

nar

-0.73

rity

-0.70

icion

-0.68

icip

-0.67

nan

-0.66

fman

-0.66

POSITIVE LOGITS

 altering

0.76

 deterior

0.76

 effected

0.75

0.74

 changed

0.74

 impacting

0.72

 differ

0.72

 alter

0.72

ãĥ£

0.71

 redesign

0.70

Activations Density 0.029%

words related to a significant and often sudden change

terms indicating significant changes or transformations

No Comments

No Known Activations