INDEX

Explanations

phrases and keywords related to safety, regulations, or compliance issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

torch.float32

Hook Point

blocks.8.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

©¶æ¥µ

-0.78

Winged

-0.74

GH

-0.72

ĸļ

-0.71

ĸļå£«

-0.69

ãĥķãĤ©

-0.69

ĪĴ

-0.68

ãĥ¯ãĥ³

-0.67

quickShipAvailable

-0.63

emale

-0.63

POSITIVE LOGITS

lees

0.74

cock

0.60

 lieu

0.57

 srfAttach

0.57

 guiActiveUnfocused

0.57

 fund

0.55

 behalf

0.55

effic

0.55

 either

0.55

imens

0.54

Activations Density 0.051%

phrases and keywords related to safety, regulations, or compliance issues

No Comments

No Known Activations