INDEX

Explanations

phrases related to descriptions of people or places

oai_token-act-pair · gpt-3.5-turbo

elements related to fictional genres and various descriptors of characters or settings within narratives

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

onyms

-0.92

mods

-0.83

soType

-0.82

osponsors

-0.79

mbuds

-0.79

ographics

-0.78

Contents

-0.78

 Cosponsors

-0.78

bits

-0.78

ayers

-0.78

POSITIVE LOGITS

 masterpiece

0.97

 enclave

0.95

 affair

0.93

 hybrid

0.87

 conco

0.82

 compound

0.82

 proposition

0.81

 creature

0.81

 sedan

0.81

 thriller

0.81

Activations Density 0.541%

phrases related to descriptions of people or places

elements related to fictional genres and various descriptors of characters or settings within narratives

No Comments

No Known Activations