INDEX

Explanations

capital letters and numbers in specific sequences

oai_token-act-pair · gpt-3.5-turbo

specific letter or number sequences and their representations within a structured data context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ification

-0.66

has

-0.63

rollers

-0.60

estate

-0.59

 Socialism

-0.59

lif

-0.58

lihood

-0.58

collection

-0.58

think

-0.55

offic

-0.55

POSITIVE LOGITS

OTO

1.09

NW

1.01

ERO

0.99

UE

0.99

-.

0.97

RL

0.97

VI

0.95

OPA

0.93

II

0.92

EGIN

0.92

Activations Density 0.110%

capital letters and numbers in specific sequences

specific letter or number sequences and their representations within a structured data context

No Comments

No Known Activations