INDEX

Explanations

information related to violent attacks causing casualties

oai_token-act-pair · gpt-3.5-turbo

references to violent incidents involving casualties

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

flush

-0.67

 motto

-0.64

ennial

-0.64

ãĤ´ãĥ³

-0.61

omial

-0.61

NetMessage

-0.60

 exception

-0.60

 COURT

-0.59

 unlocking

-0.59

 aster

-0.58

POSITIVE LOGITS

 innocent

1.14

 innoc

1.08

 bystanders

1.05

 people

1.03

 pedestrians

1.00

 civilians

0.97

 persons

0.86

 worsh

0.82

 unsuspecting

0.82

 policemen

0.81

Activations Density 0.101%

information related to violent attacks causing casualties

references to violent incidents involving casualties

No Comments

No Known Activations