INDEX

Explanations

instances of the word "aw."

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

 Luthor

-0.75

 Cerberus

-0.69

 steroids

-0.66

 Clown

-0.62

 exorc

-0.62

 Duchess

-0.61

matter

-0.61

 Painter

-0.60

 Prelude

-0.60

 suspic

-0.59

POSITIVE LOGITS

akening

1.10

aii

1.07

orld

1.06

esome

1.06

idth

1.01

ashington

1.01

arij

0.97

orks

0.96

akens

0.93

amba

0.93

Activations Density 0.009%

instances of the word "aw."

No Comments

No Known Activations