INDEX

Explanations

words related to enhancement or improvement

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

 Leilan

-0.74

HIP

-0.72

 Bullets

-0.70

 Fired

-0.68

 externalToEVAOnly

-0.68

 Funk

-0.67

 Rampage

-0.67

Nun

-0.66

LOAD

-0.65

 Shattered

-0.65

POSITIVE LOGITS

anced

1.71

ancing

1.44

ance

1.17

ancer

1.11

ancers

1.10

ances

1.01

ove

0.98

obbies

0.95

orse

0.95

urst

0.94

Activations Density 0.002%

words related to enhancement or improvement

No Comments

No Known Activations