INDEX

Explanations

the word "ills"

oai_token-act-pair · gpt-3.5-turbo

references to "vils" or similar terms related to a specific context within the document

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

uyomi

-0.77

 heartbeat

-0.73

 stake

-0.71

 stabilized

-0.70

 structured

-0.69

¥ŀ

-0.69

 subdiv

-0.68

etheless

-0.68

 lifetime

-0.67

©¶æ

-0.67

POSITIVE LOGITS

ils

1.10

ILS

0.86

uminati

0.86

aments

0.85

andr

0.85

Ãł

0.82

ung

0.81

ugi

0.80

anders

0.80

ionage

0.79

Activations Density 0.005%

the word "ills"

references to "vils" or similar terms related to a specific context within the document

No Comments

No Known Activations