INDEX

Explanations

years, locations, and names related to politics and elections

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

inates

-0.92

ctica

-0.86

ional

-0.85

uate

-0.82

inations

-0.80

orie

-0.78

ion

-0.76

rary

-0.75

ogy

-0.74

uations

-0.73

POSITIVE LOGITS

nesday

1.17

edge

0.90

fare

0.83

tip

0.83

nut

0.82

esome

0.77

ington

0.77

combe

0.76

nuts

0.76

yer

0.74

Activations Density 1.433%

years, locations, and names related to politics and elections

No Comments

No Known Activations