INDEX

Explanations

phrases related to naming or describing something

oai_token-act-pair · gpt-3.5-turbo

instances of the phrase "what [subject] calls" indicating definitions or interpretations

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

ooter

-0.72

olith

-0.67

psons

-0.65

âĨ

-0.64

etime

-0.63

usha

-0.62

————————————————

-0.62

----------------

-0.61

yrus

-0.61

atche

-0.61

POSITIVE LOGITS

XM

0.71

pires

0.64

Fra

0.64

called

0.64

 termed

0.62

interstitial

0.61

selves

0.61

 calls

0.61

illac

0.60

 sizeof

0.59

Activations Density 0.047%

phrases related to naming or describing something

instances of the phrase "what [subject] calls" indicating definitions or interpretations

No Comments

No Known Activations