INDEX

Explanations

action words related to instructions or directions

oai_token-act-pair · gpt-3.5-turbo

phrases indicating financial concerns or discussions

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

RT

-0.61

izont

-0.61

raltar

-0.61

agine

-0.60

STR

-0.60

 manifesto

-0.60

chin

-0.59

 footprint

-0.59

icipated

-0.58

irit

-0.57

POSITIVE LOGITS

 kindly

0.84

 reimburse

0.76

kson

0.73

 forbid

0.68

me

0.68

us

0.67

 theirs

0.66

 SERVICE

0.64

 hospitality

0.63

arantine

0.61

Activations Density 0.951%

action words related to instructions or directions

phrases indicating financial concerns or discussions

No Comments

No Known Activations