INDEX

Explanations

phrases related to political figures' behaviors and statements

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_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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ibrary

-0.71

ribution

-0.69

 occurs

-0.67

mination

-0.65

zb

-0.65

vidia

-0.63

ipple

-0.63

 Combine

-0.63

OOL

-0.63

igslist

-0.62

POSITIVE LOGITS

 eager

1.10

 poised

1.07

 reluctant

1.04

 nervous

1.01

 oblivious

0.99

 intent

0.98

 hesitant

0.95

 unsure

0.94

 confused

0.94

 keen

0.94

Activations Density 0.077%

phrases related to political figures' behaviors and statements

No Comments

No Known Activations