INDEX

Explanations

words related to cities, especially when discussing issues or attributes of cities

oai_token-act-pair · gpt-3.5-turbo

references to cities

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ibaba

-0.76

iasis

-0.75

selection

-0.74

raviolet

-0.70

odox

-0.68

isites

-0.66

tle

-0.66

arial

-0.65

Tro

-0.65

raid

-0.65

POSITIVE LOGITS

chool

0.95

 mayors

0.86

folk

0.85

wide

0.85

 dwell

0.79

mith

0.77

 councils

0.76

 skyline

0.76

 halls

0.76

 suburbs

0.74

Activations Density 0.020%

words related to cities, especially when discussing issues or attributes of cities

references to cities

No Comments

No Known Activations