INDEX

Explanations

words beginning with "wh," indicating inquiries or questions

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

ITAL

-0.71

ULAR

-0.67

ION

-0.66

 Grande

-0.65

 Puzzles

-0.63

WARE

-0.62

phrine

-0.60

RIC

-0.60

uated

-0.60

_>

-0.60

POSITIVE LOGITS

istle

1.37

irlwind

1.18

olly

1.17

ispers

1.15

omever

1.11

irling

1.10

isky

1.09

olen

1.06

ipp

1.06

irl

1.06

Activations Density 0.007%

words beginning with "wh," indicating inquiries or questions

No Comments

No Known Activations