INDEX

Explanations

news captions or headlines related to different global events

oai_token-act-pair · gpt-3.5-turbo

references to photographic sources or credits in the text

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

lake

-0.69

rone

-0.68

 outweigh

-0.68

auga

-0.67

anity

-0.66

icken

-0.65

roid

-0.65

oxide

-0.64

 guiActiveUn

-0.64

angel

-0.64

POSITIVE LOGITS

fecture

0.79

 millenn

0.76

Getty

0.72

Balt

0.66

issance

0.65

bably

0.64

QC

0.60

unal

0.59

 trave

0.58

kefeller

0.58

Activations Density 0.082%

news captions or headlines related to different global events

references to photographic sources or credits in the text

No Comments

No Known Activations