INDEX

Explanations

words related to press interactions in the context of the White House

oai_token-act-pair · gpt-3.5-turbo

mentions of the press or media-related terms

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

abama

-0.77

inho

-0.74

[|

-0.67

 Reconstruction

-0.67

 Valhalla

-0.67

 Scotia

-0.66

irst

-0.63

 Trayvon

-0.63

oshi

-0.63

 Haram

-0.63

POSITIVE LOGITS

urized

1.41

ured

1.02

ures

0.96

 corps

0.96

orters

0.91

room

0.89

 conferences

0.87

URE

0.86

eval

0.86

 conference

0.84

Activations Density 0.038%

words related to press interactions in the context of the White House

mentions of the press or media-related terms

No Comments

No Known Activations