INDEX

Explanations

mentions of ICE (Immigration and Customs Enforcement) activities and policies

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

hood

-0.80

 Notting

-0.78

lihood

-0.78

osity

-0.77

ipal

-0.76

acters

-0.74

amins

-0.73

wagen

-0.71

itars

-0.71

usterity

-0.69

POSITIVE LOGITS

OC

1.05

OTUS

1.02

IS

0.98

ILY

0.93

LE

0.93

CC

0.92

REAM

0.92

IG

0.92

ORN

0.92

TA

0.91

Activations Density 0.061%

mentions of ICE (Immigration and Customs Enforcement) activities and policies

No Known Activations