INDEX

Explanations

the word "each"

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Bs

-0.73

Cu

-0.68

Jer

-0.60

abad

-0.60

aucus

-0.60

stones

-0.59

XT

-0.58

etics

-0.57

Rag

-0.57

keepers

-0.57

POSITIVE LOGITS

 month

0.97

 week

0.92

 weekday

0.91

 year

0.88

 successive

0.88

"}],"

0.84

 morning

0.80

month

0.76

day

0.75

 afternoon

0.74

Activations Density 0.030%

the word "each"

No Comments

No Known Activations

the word "each"

No Comments

No Known Activations