INDEX

Explanations

phrases related to changing the course of events or specific situations

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

kit

-0.79

orthy

-0.75

©¶æ¥µ

-0.75

ores

-0.70

leeve

-0.70

Reviewer

-0.69

nice

-0.68

uania

-0.68

gur

-0.66

rake

-0.66

POSITIVE LOGITS

 events

1.13

 causation

0.97

 progression

0.96

 evolution

0.95

 succession

0.92

 traject

0.91

 development

0.90

 descent

0.88

 affairs

0.87

 progress

0.87

Activations Density 0.171%

phrases related to changing the course of events or specific situations

No Comments

No Known Activations