INDEX

Explanations

pronouns referring to abstract concepts such as actions or beliefs

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

 Spoon

-0.70

 Nano

-0.70

 Aluminum

-0.69

 Pirates

-0.66

 Sunshine

-0.66

adelphia

-0.65

Topics

-0.61

 Firefly

-0.61

holding

-0.59

 Kang

-0.59

POSITIVE LOGITS

self

0.99

 oneself

0.97

 ourselves

0.83

 altogether

0.82

alian

0.81

atic

0.80

 accordingly

0.79

atical

0.78

 uncond

0.77

iner

0.75

Activations Density 0.154%

pronouns referring to abstract concepts such as actions or beliefs

No Comments

No Known Activations