INDEX

Explanations

sequences of repeated letters

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

ĪĴ

-0.77

 pregn

-0.70

BIT

-0.65

 Marketable

-0.58

 pity

-0.54

 burg

-0.54

Legend

-0.53

 omission

-0.53

INAL

-0.53

ISA

-0.53

POSITIVE LOGITS

lein

0.92

heit

0.84

uddin

0.84

arella

0.82

opot

0.80

eus

0.76

enne

0.76

iland

0.75

hesive

0.73

aney

0.70

Activations Density 0.014%

sequences of repeated letters

No Comments

No Known Activations