INDEX

Explanations

details in a story related to character choices and circumstances

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

etimes

-0.67

âĸ¬

-0.63

mosp

-0.63

 Accountability

-0.58

 oath

-0.57

lore

-0.56

ribute

-0.55

isman

-0.55

 vengeance

-0.54

ivities

-0.54

POSITIVE LOGITS

 onwards

0.86

0.74

 onward

0.70

above

0.69

 below

0.69

cca

0.67

0.66

xxxx

0.66

âĢ²

0.66

below

0.66

Activations Density 0.117%

details in a story related to character choices and circumstances

No Comments

No Known Activations