INDEX

Explanations

phrases related to war and conflict, specifically mentioning attacks, clashes, and military actions

oai_token-act-pair · gpt-3.5-turbo

events related to violence and conflict

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

 designing

-0.82

 royalty

-0.82

 commissions

-0.81

ennial

-0.74

 maximizing

-0.74

 brainstorm

-0.72

 envision

-0.70

 audition

-0.70

yrights

-0.69

 Pebble

-0.69

POSITIVE LOGITS

 Reuters

0.85

 Serbian

0.83

UTERS

0.82

 Rohingya

0.80

 Indones

0.80

 Nato

0.80

 bombed

0.80

 deadliest

0.79

BBC

0.79

Rw

0.78

Activations Density 0.735%

phrases related to war and conflict, specifically mentioning attacks, clashes, and military actions

events related to violence and conflict

No Comments

No Known Activations