INDEX

Explanations

phrases indicating measurements or rates, particularly those involving time or quantities

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

enegger

-0.78

 Wilhelm

-0.75

FORE

-0.74

baugh

-0.74

ARDIS

-0.73

hess

-0.73

ãĤ¶

-0.72

 Sloan

-0.72

soDeliveryDate

-0.71

BACK

-0.71

POSITIVE LOGITS

 capita

1.13

 gallon

0.91

imeter

0.89

 hour

0.85

 increment

0.85

missible

0.84

imeters

0.79

missive

0.77

tains

0.76

 quarter

0.75

Activations Density 0.016%

phrases indicating measurements or rates, particularly those involving time or quantities

No Comments

No Known Activations