INDEX

Explanations

phrases related to legal or political 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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

iva

-0.71

 irrespective

-0.71

acular

-0.68

 realised

-0.66

 sacrific

-0.64

 sights

-0.59

 tremend

-0.59

iors

-0.59

 pract

-0.59

BILITIES

-0.59

POSITIVE LOGITS

 Scrib

0.88

 Wikileaks

0.87

 WikiLeaks

0.80

ebin

0.80

 FOIA

0.77

 February

0.76

 redacted

0.75

 Deadline

0.74

 March

0.71

 archive

0.71

Activations Density 0.218%

phrases related to legal or political matters

No Comments

No Known Activations