INDEX

Explanations

phrases indicating completeness or detail in descriptions

oai_token-act-pair · gpt-3.5-turbo

phrases that describe something being completed or accompanied by specific attributes or elements

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

thren

-0.72

ungle

-0.71

sts

-0.70

borgh

-0.70

orah

-0.69

nery

-0.69

ights

-0.67

lisher

-0.67

ivas

-0.67

testing

-0.66

POSITIVE LOGITS

stood

1.03

 regard

0.81

draw

0.79

 bells

0.79

 nails

0.78

 extras

0.75

drawn

0.73

 impunity

0.72

 jewels

0.71

ttes

0.71

Activations Density 0.149%

phrases indicating completeness or detail in descriptions

phrases that describe something being completed or accompanied by specific attributes or elements

No Comments

No Known Activations