INDEX

Explanations

questions starting with "How did" or "Why did"

oai_token-act-pair · gpt-3.5-turbo

questions starting with "How" or "Did."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

houses

-0.74

heter

-0.72

washer

-0.71

 Methods

-0.71

thur

-0.70

rooms

-0.70

arters

-0.69

room

-0.68

atten

-0.68

north

-0.68

POSITIVE LOGITS

actic

1.01

iosyncr

0.82

netflix

0.77

nt

0.68

IER

0.68

 originate

0.68

ĸļ

0.67

not

0.66

riks

0.65

 undergo

0.65

Activations Density 0.042%

questions starting with "How did" or "Why did"

questions starting with "How" or "Did."

No Comments

No Known Activations