INDEX

Explanations

terms related to identity and social categorization

oai_token-act-pair · gpt-4o-mini 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 heel

-0.67

 straw

-0.65

 covering

-0.64

 grounds

-0.61

 purse

-0.60

 overlooking

-0.60

hue

-0.60

wig

-0.60

 hearts

-0.59

 Springer

-0.58

POSITIVE LOGITS

alist

1.39

ally

1.34

ual

1.25

alities

1.21

al

1.20

alis

1.20

ality

1.18

inarily

1.17

amental

1.15

ional

1.15

Activations Density 0.152%

terms related to identity and social categorization

No Comments

No Known Activations