INDEX

Explanations

references to legal cases and political figures

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

uria

-0.89

cedes

-0.82

thia

-0.80

ourning

-0.80

riel

-0.79

duty

-0.78

NetMessage

-0.76

verbs

-0.76

ocrates

-0.76

requires

-0.72

POSITIVE LOGITS

 debacle

1.05

 playbook

1.04

era

1.00

 universe

0.96

 fiasco

0.96

 bandwagon

0.96

 saga

0.95

 portion

0.95

 Foundation

0.94

 Institute

0.93

Activations Density 1.959%

references to legal cases and political figures

No Comments

No Known Activations