INDEX

Explanations

phrases related to governmental policies and different societal perspectives

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

 Anyway

-0.74

 Bound

-0.69

 Apart

-0.68

 Delicious

-0.66

Okay

-0.63

Alright

-0.62

 Trop

-0.62

 Anything

-0.61

PokÃ©

-0.61

Rid

-0.60

POSITIVE LOGITS

 specialize

1.11

oping

1.04

 weren

1.01

 benefited

1.00

 participated

1.00

oped

0.97

 dare

0.97

 aren

0.95

 reside

0.95

 aspire

0.95

Activations Density 0.550%

phrases related to governmental policies and different societal perspectives

No Comments

No Known Activations