INDEX

Explanations

patterns related to chronological actions or sequences

oai_token-act-pair · gpt-3.5-turbo

concepts related to progression or completion in a sequential manner

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

?)

-0.76

-)

-0.68

hack

-0.67

?]

-0.66

panel

-0.63

%]

-0.62

arten

-0.61

~~~~

-0.60

?)

-0.59

 inaction

-0.58

POSITIVE LOGITS

 respectively

0.84

 effortlessly

0.78

 spew

0.74

 culmin

0.70

 neatly

0.68

 obliter

0.68

 brutally

0.68

 perched

0.67

 seamlessly

0.66

 sandwic

0.65

Activations Density 1.251%

patterns related to chronological actions or sequences

concepts related to progression or completion in a sequential manner

No Comments

No Known Activations