INDEX

Explanations

mentions of specific news networks and their associated programs or anchors

oai_token-act-pair · gpt-3.5-turbo

references to specific television news networks and their personalities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

jar

-0.83

 Strongh

-0.82

ABV

-0.79

idas

-0.78

shall

-0.74

abiding

-0.74

chest

-0.74

borgh

-0.73

nants

-0.72

range

-0.71

POSITIVE LOGITS

 reporter

1.42

 correspondent

1.36

 commentator

1.36

 columnist

1.34

 host

1.32

 anchor

1.31

 pund

1.26

 interviewer

1.23

 journalist

1.22

 contributor

1.22

Activations Density 0.143%

mentions of specific news networks and their associated programs or anchors

references to specific television news networks and their personalities

No Comments

No Known Activations