INDEX

Explanations

locations or geographic regions

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

thood

-0.90

urated

-0.86

nl

-0.85

 arrang

-0.79

selage

-0.79

emort

-0.77

acles

-0.77

earance

-0.77

manship

-0.74

onymous

-0.74

POSITIVE LOGITS

 Forge

1.00

 Regional

0.85

 Fever

0.82

 Valley

0.82

 Unified

0.79

 Grizzlies

0.76

zx

0.75

 Stream

0.74

 Region

0.73

 Parkway

0.72

Activations Density 0.096%

locations or geographic regions

No Comments

No Known Activations