INDEX

Explanations

words associated with confusion or instability

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

 refined

-0.65

ngth

-0.64

catentry

-0.62

ner

-0.58

 diction

-0.55

Dial

-0.54

 gifted

-0.54

 libel

-0.53

chell

-0.53

 towed

-0.53

POSITIVE LOGITS

ither

0.89

rift

0.85

asa

0.78

oros

0.76

ewater

0.76

acia

0.72

abus

0.72

alys

0.71

oon

0.71

ike

0.69

Activations Density 0.031%

words associated with confusion or instability

No Comments

No Known Activations