INDEX

Explanations

information related to destructive events like fires, bombings, closures, evacuations, and robberies in various locations

oai_token-act-pair · gpt-3.5-turbo

references to destruction or damage to properties and infrastructure

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

verse

-0.74

Austral

-0.73

 inexper

-0.69

=-=-=-=-

-0.69

 impulses

-0.69

 versatility

-0.65

vich

-0.65

who

-0.65

 theorem

-0.64

rities

-0.64

POSITIVE LOGITS

 flooded

1.11

 reopened

1.06

 evacuated

1.04

 overrun

1.03

ģ«

1.02

 congest

0.98

 inund

0.98

 habitable

0.97

 seism

0.97

 emptied

0.97

Activations Density 0.490%

information related to destructive events like fires, bombings, closures, evacuations, and robberies in various locations

references to destruction or damage to properties and infrastructure

No Comments

No Known Activations