INDEX

Explanations

instances of the word "round" followed by a numeric value

oai_token-act-pair · gpt-3.5-turbo

references to rounds of negotiations or discussions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 sqor

-0.78

esse

-0.77

sense

-0.77

reci

-0.73

iverpool

-0.68

ply

-0.65

Planet

-0.64

ilk

-0.64

plement

-0.64

ghan

-0.64

POSITIVE LOGITS

 applause

1.12

 layoffs

0.89

 hostilities

0.84

 chemotherapy

0.83

 negotiations

0.80

 hearings

0.78

 eviction

0.78

 reckoning

0.75

 tests

0.74

 playoffs

0.72

Activations Density 0.097%

instances of the word "round" followed by a numeric value

references to rounds of negotiations or discussions

No Comments

No Known Activations