INDEX

Explanations

instances of the word "almost" followed by a number

oai_token-act-pair · gpt-3.5-turbo

references to the word "almost" in various contexts

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

oran

-0.79

agate

-0.79

oris

-0.73

RTX

-0.71

Ds

-0.70

eria

-0.66

alam

-0.65

oÄŁ

-0.65

Ey

-0.64

erion

-0.63

POSITIVE LOGITS

etheless

0.79

 certainly

0.74

 exclusively

0.73

 identical

0.71

stress

0.69

 mundane

0.67

zero

0.66

 ident

0.63

 electr

0.63

untarily

0.63

Activations Density 0.032%

instances of the word "almost" followed by a number

references to the word "almost" in various contexts

No Comments

No Known Activations