INDEX

Explanations

specific locations or settings

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "here," indicating a focus on local references or 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

eer

-0.65

 Thrones

-0.61

 Decay

-0.59

IRC

-0.58

 Spending

-0.57

 Kaiser

-0.54

 Meter

-0.54

CHAT

-0.54

ahime

-0.54

 Shards

-0.53

POSITIVE LOGITS

tics

1.91

tical

1.82

abouts

1.64

tic

1.51

to

0.97

with

0.96

upon

0.88

ina

0.83

after

0.79

from

0.76

Activations Density 0.057%

specific locations or settings

occurrences of the word "here," indicating a focus on local references or context

No Comments

No Known Activations