INDEX

Explanations

the word "hostages" or related terms

oai_token-act-pair · gpt-3.5-turbo

references to hostages and negotiations related to hostage situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

agate

-0.76

 Arbor

-0.72

Rot

-0.71

mor

-0.71

River

-0.71

Measure

-0.68

ycle

-0.67

alia

-0.67

 unfocusedRange

-0.66

bm

-0.64

POSITIVE LOGITS

 hostages

2.92

 hostage

1.97

 negotiator

1.76

 negotiators

1.76

 captives

1.67

 bystanders

1.31

 captive

1.29

 standoff

1.13

 bystand

1.10

goers

1.10

Activations Density 0.049%

the word "hostages" or related terms

references to hostages and negotiations related to hostage situations

No Comments

No Known Activations