INDEX

Explanations

references to locations or entities named "Main."

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

IBLE

-0.70

uracy

-0.67

âķĲâķĲ

-0.61

terday

-0.61

OPE

-0.60

hovah

-0.59

 perfection

-0.59

 tender

-0.59

ossession

-0.59

orthy

-0.58

POSITIVE LOGITS

tenance

1.64

stay

1.20

stream

1.16

deck

1.07

lander

1.03

land

0.97

yu

0.95

stad

0.93

frame

0.92

tan

0.89

Activations Density 0.010%

references to locations or entities named "Main."

No Comments

No Known Activations