INDEX

Explanations

profanity and expletives

oai_token-act-pair · gpt-3.5-turbo

expressions of frustration or disbelief

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

Joy

-0.75

atures

-0.67

FN

-0.66

 obser

-0.65

iva

-0.63

Mos

-0.62

RNA

-0.61

omore

-0.61

Mesh

-0.61

izont

-0.61

POSITIVE LOGITS

 happened

1.02

 happens

0.79

 else

0.79

 transpired

0.79

ARE

0.77

 ensued

0.74

 THEY

0.72

!?"

0.72

chu

0.72

?!

0.71

Activations Density 0.059%

profanity and expletives

expressions of frustration or disbelief

No Comments

No Known Activations