INDEX

Explanations

words related to items or concepts that are of a lower value or rank

oai_token-act-pair · gpt-3.5-turbo

references to the word "low" 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 @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

tnc

-0.94

andise

-0.81

itutional

-0.69

ADRA

-0.67

andum

-0.67

 Orient

-0.67

illon

-0.66

ilage

-0.65

ophers

-0.65

 Pengu

-0.64

POSITIVE LOGITS

ered

0.91

est

0.90

enthal

0.88

ball

0.83

 hanging

0.81

(<

0.78

 enough

0.77

ppy

0.77

 pitched

0.76

down

0.76

Activations Density 0.039%

words related to items or concepts that are of a lower value or rank

references to the word "low" in various contexts

No Comments

No Known Activations