INDEX

Explanations

words related to systems and structures

oai_token-act-pair · gpt-3.5-turbo

references to various systems, highlighting their roles or impacts in different contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

)]

-0.76

tein

-0.75

ishable

-0.73

 earnest

-0.73

ãĤ¤ãĥĪ

-0.71

ãĤ¨ãĥ«

-0.67

IGHTS

-0.66

æĢ

-0.66

Fil

-0.64

Bone

-0.64

POSITIVE LOGITS

 system

0.96

atics

0.93

 systems

0.92

 integ

0.81

 lawy

0.74

ycle

0.73

 sshd

0.73

 disg

0.72

 outage

0.72

 hierarchy

0.71

Activations Density 0.041%

words related to systems and structures

references to various systems, highlighting their roles or impacts in different contexts

No Comments

No Known Activations