INDEX

Explanations

the word "thoughts"

oai_token-act-pair · gpt-3.5-turbo

references to personal thoughts or opinions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

ALL

-0.77

 Mamm

-0.69

Claim

-0.69

ARDS

-0.61

toe

-0.61

Ann

-0.58

 Breach

-0.57

rake

-0.56

ards

-0.56

announced

-0.55

POSITIVE LOGITS

 aloud

0.88

fulness

0.87

cience

0.85

ileaks

0.84

ynthesis

0.79

mith

0.79

umar

0.77

 thoughts

0.77

warts

0.75

matter

0.73

Activations Density 0.017%

the word "thoughts"

references to personal thoughts or opinions

No Comments

No Known Activations