INDEX

Explanations

terms related to political parties, especially the Democratic Party

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

icles

-0.95

paio

-0.80

ooter

-0.79

obar

-0.76

icle

-0.75

ordon

-0.75

mares

-0.74

renheit

-0.73

ooters

-0.73

asus

-0.73

POSITIVE LOGITS

 Party

1.20

 strongh

0.94

 stronghold

0.94

 caucuses

0.84

 nominee

0.83

 frontrunner

0.80

 caucus

0.78

 senator

0.78

 Caucus

0.78

 operative

0.77

Activations Density 0.467%

terms related to political parties, especially the Democratic Party

No Comments

No Known Activations