INDEX

Explanations

words related to governmental and political topics

oai_token-act-pair · gpt-3.5-turbo

references to consequences of actions or events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Rescue

-0.79

ptives

-0.74

theless

-0.73

ventus

-0.66

gaard

-0.66

anova

-0.65

 Locations

-0.64

hops

-0.63

nces

-0.62

OGR

-0.62

POSITIVE LOGITS

icter

0.89

 timeframe

0.73

liest

0.72

 fielder

0.66

 unborn

0.64

 portion

0.63

 outwe

0.61

 ecosystem

0.60

aeper

0.59

ename

0.59

Activations Density 3.356%

words related to governmental and political topics

references to consequences of actions or events

No Comments

No Known Activations