INDEX

Explanations

phrases related to specific organizations or events, potentially related to news or political content

oai_token-act-pair · gpt-3.5-turbo

phrases or constructions that include the word "with."

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

EMBER

-0.65

Polit

-0.64

BF

-0.61

Generally

-0.61

UFC

-0.61

polit

-0.61

é¾

-0.60

Merit

-0.59

NN

-0.59

 encount

-0.59

POSITIVE LOGITS

stood

1.46

draw

1.35

drawn

1.33

 regards

1.06

 regard

1.05

 apologies

1.05

holding

0.98

 caveats

0.97

 exception

0.90

 whom

0.90

Activations Density 0.104%

phrases related to specific organizations or events, potentially related to news or political content

phrases or constructions that include the word "with."

No Comments

No Known Activations