INDEX

Explanations

references to government entities or official state matters

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

poons

-0.94

eger

-0.77

aurus

-0.76

ogle

-0.74

edin

-0.73

 Aires

-0.71

sites

-0.69

 Brilliant

-0.68

Maw

-0.68

icago

-0.66

POSITIVE LOGITS

 legislatures

1.19

 legislature

1.05

wide

1.04

 governments

0.91

 Legislature

0.89

 trooper

0.89

 troopers

0.87

 legislators

0.87

 governors

0.87

 legislator

0.87

Activations Density 0.436%

references to government entities or official state matters

No Comments

No Known Activations