INDEX

Explanations

words related to indications or signals of events or conditions

oai_token-act-pair · gpt-3.5-turbo

phrases regarding signs of change or condition

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ahime

-0.74

ISH

-0.74

ategory

-0.73

uble

-0.72

halla

-0.69

UTE

-0.69

adelphia

-0.66

OAD

-0.64

aughs

-0.64

[|

-0.62

POSITIVE LOGITS

 signs

0.96

 Signs

0.95

achy

0.77

 indicating

0.73

igmatic

0.70

 waning

0.69

ifact

0.69

posts

0.69

 deterior

0.69

 deteriorated

0.67

Activations Density 0.027%

words related to indications or signals of events or conditions

phrases regarding signs of change or condition

No Comments

No Known Activations