INDEX

Explanations

solutions or actions outlined in a text

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

Negative Logits

 Referred

-0.77

 Conference

-0.72

 conferences

-0.70

 Cosponsors

-0.66

 Expansion

-0.66

 conference

-0.64

NAS

-0.61

estern

-0.60

 Contracts

-0.60

 internationally

-0.59

POSITIVE LOGITS

 suffice

1.44

 dissu

0.89

 overwhelm

0.88

 alleviate

0.87

SPONSORED

0.86

help

0.85

 cure

0.84

 overpower

0.82

wered

0.82

 detract

0.81

Activations Density 0.422%

solutions or actions outlined in a text

No Known Activations