INDEX

Explanations

phrases indicating a shift or change in focus

oai_token-act-pair · gpt-3.5-turbo

coordinating conjunctions that signify a shift in focus or agenda

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

one

-0.68

geries

-0.67

MK

-0.67

ibles

-0.67

wat

-0.67

mods

-0.66

anks

-0.66

among

-0.66

four

-0.66

anka

-0.65

POSITIVE LOGITS

 instead

1.38

 Instead

1.32

Instead

1.30

 replaced

1.30

 concentrate

1.21

 opting

1.15

instead

1.15

 focus

1.08

 replaces

1.02

 substituted

1.00

Activations Density 0.211%

phrases indicating a shift or change in focus

coordinating conjunctions that signify a shift in focus or agenda

No Comments

No Known Activations