INDEX

Explanations

terms related to controversial, scandalous, and high-profile events or cases

oai_token-act-pair · gpt-3.5-turbo

references to scandals and significant legal cases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

ordinate

-0.76

atin

-0.73

yrics

-0.72

VEL

-0.69

ummer

-0.68

ccording

-0.67

usable

-0.67

background

-0.67

vae

-0.66

ailable

-0.65

POSITIVE LOGITS

 debacle

1.01

 fiasco

0.92

naires

0.87

 spree

0.84

naire

0.80

undrum

0.76

 saga

0.75

 trilogy

0.74

AMC

0.73

ilogy

0.71

Activations Density 0.294%

terms related to controversial, scandalous, and high-profile events or cases

references to scandals and significant legal cases

No Comments

No Known Activations