INDEX

Explanations

references to political figures and events

oai_token-act-pair · gpt-3.5-turbo

numerical values and identifiers, particularly related to structured data or lists

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Cyrus

-0.70

heit

-0.67

Quantity

-0.65

 tranquil

-0.65

tnc

-0.62

 secrecy

-0.62

ĸļ

-0.60

Kus

-0.60

 Ellis

-0.59

 privacy

-0.58

POSITIVE LOGITS

GHz

0.79

ateral

0.77

idon

0.76

peat

0.75

teness

0.75

iev

0.74

thodox

0.72

drm

0.72

alion

0.70

ilver

0.69

Activations Density 0.240%

references to political figures and events

numerical values and identifiers, particularly related to structured data or lists

No Comments

No Known Activations