INDEX

Explanations

terms related to law enforcement and legal actions

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

actions related to implementing bans and restrictions

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

ebus

-0.81

Sov

-0.72

rious

-0.70

rouse

-0.66

united

-0.66

 Dynamics

-0.65

gow

-0.65

swick

-0.65

lycer

-0.64

orthy

-0.62

POSITIVE LOGITS

 altogether

1.21

 unnecessary

0.98

 offending

0.94

 entirely

0.94

 outright

0.90

 abruptly

0.89

 unwanted

0.88

 bothering

0.86

 disbelief

0.82

 inhib

0.81

Activations Density 4.360%

terms related to law enforcement and legal actions

actions related to implementing bans and restrictions

No Comments

No Known Activations