INDEX

Explanations

phrases related to politics, legislation, and international affairs

oai_token-act-pair · gpt-3.5-turbo

terms related to classification and categorization, particularly in a societal or systemic context

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

 Leilan

-0.50

Io

-0.50

 respectively

-0.50

$$$$

-0.45

 Mara

-0.45

spin

-0.44

 goodbye

-0.42

Abu

-0.42

 Fitz

-0.41

ä»£

-0.40

POSITIVE LOGITS

issions

0.60

ibilities

0.59

ruct

0.58

Ĩ

0.57

ential

0.57

atures

0.57

itives

0.56

itionally

0.56

ensed

0.56

ework

0.55

Activations Density 0.389%

phrases related to politics, legislation, and international affairs

terms related to classification and categorization, particularly in a societal or systemic context

No Comments

No Known Activations