INDEX

Explanations

words related to specific names and locations, possibly related to news articles or reports

oai_token-act-pair · gpt-3.5-turbo

names and surnames, particularly those that are repeated in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

osate

-0.76

ISC

-0.76

imil

-0.74

NES

-0.72

ilt

-0.72

orb

-0.72

 antid

-0.71

pmwiki

-0.70

itars

-0.69

ennis

-0.69

POSITIVE LOGITS

lings

0.77

 lang

0.74

rouch

0.69

wolves

0.68

mast

0.66

$$$$

0.64

mans

0.63

angler

0.63

rained

0.62

Compan

0.61

Activations Density 0.017%

words related to specific names and locations, possibly related to news articles or reports

names and surnames, particularly those that are repeated in the text

No Comments

No Known Activations