INDEX

Explanations

word related to skepticism towards scientific and medical establishments

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

 Compass

-0.83

canon

-0.82

 Trace

-0.81

TPPStreamerBot

-0.80

 Recorded

-0.75

 Finder

-0.74

byter

-0.73

availability

-0.69

 Skip

-0.69

inguished

-0.68

POSITIVE LOGITS

 scapego

1.20

 impunity

1.11

 gull

1.11

 pretext

1.10

 extortion

1.08

 bogus

1.06

 blackmail

1.06

 phony

1.02

 bribes

1.02

 exploiting

1.00

Activations Density 3.756%

word related to skepticism towards scientific and medical establishments

No Comments

No Known Activations