INDEX

Explanations

textual elements related to data representation or processing

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

 camping

-0.78

izer

-0.76

 Boone

-0.76

 planner

-0.74

 billboard

-0.74

 recorder

-0.73

clinton

-0.72

 Hilton

-0.72

 patio

-0.71

conservancy

-0.70

POSITIVE LOGITS

Â

1.93

Â±

1.37

Â½

1.31

Â²

1.25

Â¢

1.24

Â£

1.23

Å

1.17

Ã

1.15

Â«

1.15

Â°

1.14

Activations Density 0.009%

textual elements related to data representation or processing

No Comments

No Known Activations