INDEX

Explanations

mentions of security-related terms or concepts within geopolitical contexts

oai_token-act-pair · gpt-3.5-turbo

terms and phrases related to security issues

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

igs

-0.79

xt

-0.77

igible

-0.72

arget

-0.70

ICA

-0.69

zos

-0.67

alogy

-0.67

ascar

-0.67

ergy

-0.67

alog

-0.67

POSITIVE LOGITS

 advis

0.93

 Adviser

0.90

 Advisor

0.87

 clearance

0.87

 adviser

0.85

policy

0.84

 personnel

0.84

 guards

0.84

 advisor

0.83

 Forces

0.82

Activations Density 0.023%

mentions of security-related terms or concepts within geopolitical contexts

terms and phrases related to security issues

No Comments

No Known Activations