INDEX

Explanations

references to legal actions and disputes involving prominent figures or institutions

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

holes

-0.74

ordes

-0.66

okers

-0.64

cil

-0.64

adle

-0.63

oute

-0.63

obin

-0.60

lectic

-0.59

aunders

-0.59

/#

-0.58

POSITIVE LOGITS

 week

1.22

 month

1.20

 year

1.12

 ditch

1.00

 night

1.00

 rites

0.92

 Month

0.90

 semester

0.89

 summer

0.89

 weekend

0.89

Activations Density 0.702%

references to legal actions and disputes involving prominent figures or institutions

No Comments

No Known Activations