INDEX

Explanations

excerpts related to personal interactions and relationships

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

 Samar

-0.81

 Skydragon

-0.75

 Gmail

-0.73

ZIP

-0.66

BBC

-0.64

 universities

-0.63

MUS

-0.63

 Democracy

-0.62

 fragmentation

-0.62

 pressures

-0.62

POSITIVE LOGITS

ve

1.11

sure

1.10

felt

1.09

ved

1.08

shall

1.03

ski

1.02

re

1.00

vest

1.00

Activations Density 0.628%

excerpts related to personal interactions and relationships

No Comments

No Known Activations