INDEX

Explanations

mentions of academic subjects or concepts, particularly mathematics and science

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

Sov

-0.79

ktop

-0.70

lease

-0.65

views

-0.65

Dialogue

-0.61

hold

-0.60

 Marcos

-0.59

 flesh

-0.59

 Border

-0.58

 Luxem

-0.58

POSITIVE LOGITS

matical

1.18

hematic

1.12

hemat

1.02

ilda

1.01

ieu

1.01

ilde

0.98

ias

0.92

 equations

0.90

ians

0.86

gebra

0.86

Activations Density 0.027%

mentions of academic subjects or concepts, particularly mathematics and science

No Comments

No Known Activations