INDEX

Explanations

words related to surveillance, politics, and urban environments

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

aceutical

-0.76

Constructed

-0.70

enegger

-0.69

UTION

-0.64

ersion

-0.63

ometers

-0.63

uality

-0.63

ãĥ¼ãĥ³

-0.62

-0.61

POSITIVE LOGITS

reth

1.07

lette

1.05

vernment

1.02

rier

0.99

hou

0.98

rer

0.93

bles

0.92

levard

0.92

pees

0.92

rown

0.90

Activations Density 0.794%

words related to surveillance, politics, and urban environments

No Comments

No Known Activations