INDEX

Explanations

names of specific individuals, especially female names like "She", "Sophie", and "Sarah", and relevant personal pronouns like "her", "she", and "herself"

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

ype

-0.83

kefeller

-0.82

ypes

-0.80

undo

-0.76

ozy

-0.76

vernment

-0.72

iago

-0.70

raltar

-0.69

ornia

-0.69

aleb

-0.68

POSITIVE LOGITS

 herself

1.33

pher

1.26

athed

1.10

 Anne

1.03

athing

0.99

pard

0.96

 vagina

0.95

 husband

0.94

 breasts

0.93

 Marie

0.93

Activations Density 1.504%

names of specific individuals, especially female names like "She", "Sophie", and "Sarah", and relevant personal pronouns like "her", "she", and "herself"

No Comments

No Known Activations