INDEX

Explanations

terms related to terrorist activities and attacks

oai_token-act-pair · gpt-3.5-turbo

references to terrorist attacks and related events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

é¾įå

-0.82

 Input

-0.81

Opt

-0.80

 CONTR

-0.79

 Bench

-0.74

éĸ

-0.72

 Adjusted

-0.71

River

-0.70

 Construct

-0.69

 Vale

-0.69

POSITIVE LOGITS

 bombings

1.24

 terrorists

1.18

 terror

1.18

 terrorist

1.17

 bombing

1.13

 terrorism

1.06

 gunman

1.06

 bombers

1.02

 jihadists

1.01

 shootings

1.00

Activations Density 0.238%

terms related to terrorist activities and attacks

references to terrorist attacks and related events

No Comments

No Known Activations