INDEX

Explanations

names, particularly the name "Tol" with varying activations indicating different contexts

oai_token-act-pair · gpt-3.5-turbo

references to a specific influential figure or their works

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

VD

-0.66

 Towns

-0.64

advertisement

-0.60

 abras

-0.60

 embedded

-0.59

dq

-0.59

CD

-0.59

 vaccines

-0.58

key

-0.57

 wedge

-0.57

POSITIVE LOGITS

teness

0.98

folios

0.89

mination

0.86

Ðº

0.83

ledged

0.82

theless

0.81

urations

0.81

uden

0.79

Ñı

0.78

mie

0.75

Activations Density 0.055%

names, particularly the name "Tol" with varying activations indicating different contexts

references to a specific influential figure or their works

No Comments

No Known Activations