INDEX

Explanations

phrases related to law enforcement and government investigations

oai_token-act-pair · gpt-3.5-turbo

terms related to governmental and institutional investigations

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rooms

-0.76

oning

-0.75

holes

-0.75

orously

-0.74

oÄŁ

-0.74

oned

-0.74

lain

-0.73

onies

-0.72

worn

-0.72

lessly

-0.71

POSITIVE LOGITS

 Affairs

1.33

 Intelligence

1.15

 Services

1.11

 Forces

1.06

 Development

1.06

 Relations

1.04

 Authority

1.03

Aid

1.00

 Branch

0.98

 Advisor

0.98

Activations Density 0.152%

phrases related to law enforcement and government investigations

terms related to governmental and institutional investigations

No Comments

No Known Activations