INDEX

Explanations

phrases related to information states and logic

oai_token-act-pair · gpt-3.5-turbo

references to binary states or distinctions

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

Ire

-0.60

YD

-0.57

aban

-0.57

HCR

-0.56

VERTISEMENT

-0.56

illi

-0.56

soType

-0.55

 Leban

-0.54

umni

-0.53

tten

-0.53

POSITIVE LOGITS

 versus

0.60

 programmers

0.59

 hypothetical

0.59

 chess

0.58

 mathematic

0.57

XY

0.56

\(

0.56

 pawn

0.54

 perceptual

0.54

 randomly

0.53

Activations Density 2.342%

phrases related to information states and logic

references to binary states or distinctions

No Comments

No Known Activations