INDEX

Explanations

Specific "<start>" and "<end>" tokens

oai_token-act-pair · gpt-3.5-turbo

segments related to events, decisions, or timelines, particularly in a legal or procedural context

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

senal

-0.75

 eleph

-0.71

 exting

-0.69

 undermin

-0.68

 tremend

-0.67

 subur

-0.65

aditional

-0.64

ThumbnailImage

-0.64

aturdays

-0.61

ometimes

-0.59

POSITIVE LOGITS

CVE

0.66

idable

0.65

jad

0.62

wm

0.61

gex

0.61

↵

0.60

Domin

0.59

Trend

0.59

AppData

0.59

________________________________________________________________

0.58

Activations Density 0.566%

Specific "<start>" and "<end>" tokens

segments related to events, decisions, or timelines, particularly in a legal or procedural context

No Comments

No Known Activations