INDEX

Explanations

words related to deep significance or impact

oai_token-act-pair · gpt-3.5-turbo

references to significant or impactful concepts

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

©¶æ

-0.87

phis

-0.73

rera

-0.69

oaded

-0.65

 Garrison

-0.63

avers

-0.62

annis

-0.62

hops

-0.62

 Fighters

-0.60

matched

-0.60

POSITIVE LOGITS

est

0.87

ly

0.79

ness

0.78

 philosophical

0.73

 sadness

0.71

 impact

0.69

itudinal

0.69

edIn

0.69

 amounts

0.65

ulty

0.65

Activations Density 0.013%

words related to deep significance or impact

references to significant or impactful concepts

No Comments

No Known Activations