INDEX

Explanations

names of individuals involved in legal or political actions

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

enhagen

-0.75

undo

-0.70

SDL

-0.66

solete

-0.65

raltar

-0.65

 rall

-0.64

jandro

-0.64

odder

-0.64

ighth

-0.63

Players

-0.63

POSITIVE LOGITS

 herself

2.15

 husband

1.21

 Louise

1.13

her

1.11

 hijab

1.10

 maid

1.09

 breasts

1.08

 pregnant

1.08

she

1.06

 bikini

1.06

Activations Density 2.478%

names of individuals involved in legal or political actions

No Comments

No Known Activations