INDEX

Explanations

news article captions or headlines

oai_token-act-pair · gpt-3.5-turbo

phrases related to visual media or photography

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

nesia

-0.71

eele

-0.70

nir

-0.69

lite

-0.68

roth

-0.67

sole

-0.66

letter

-0.66

mare

-0.66

abase

-0.66

STATE

-0.64

POSITIVE LOGITS

 Photos

0.79

VERTISEMENT

0.77

 Deaths

0.77

ideos

0.72

 Faces

0.68

natureconservancy

0.68

 Celeb

0.66

etheus

0.66

ahl

0.66

 Moments

0.65

Activations Density 0.045%

news article captions or headlines

phrases related to visual media or photography

No Comments

No Known Activations